Chapter 19. Traumatic Brain Injury

Traumatic brain injury (TBI) is a leading cause of death and disability among children and young adults.1,2 Unintentional injuries (ie, car accidents and falls) are the leading cause of death for children 1 to 14 years of age.3 Among such injuries, TBI is a leading cause of injury-related morbidity and mortality.4 Furthermore, despite modern automobile design and injury prevention campaigns, important causes of TBI have increased in recent years.5 In addition to unintentional injuries, child abuse remains a significant problem and constitutes the leading cause of serious head injury in infants.6 With nearly half a million children affected each year, TBI is a serious public health problem.7

Traditionally, TBI severity has been defined using the Glasgow Coma Scale (GCS). The GCS, shown in Table 19-1, was developed in order to standardize the neurological assessment of adult patients with traumatic brain injury.8 It was specifically designed to be easily performed based upon clinical data and to have a low rate of inter-observer variability. Despite its limitations when applied to children,9,10 the GCS is widely used for the initial assessment and for monitoring progress of pediatric TBI. A pediatric version of the GCS also seems to be a reliable tool for predicting the need for acute intervention in preverbal children with TBI.11

The GCS score is determined by adding the values for eye opening, verbal response, and motor response. Possible values range from 3 to 15. Note that this scale rates the best response only. In patients who are intubated, in whom assessment of best verbal response cannot be performed, notation of this is made in the GCS score by adding a "T" to the end of the score. In patients who are intubated, the best possible score would therefore be 11T. For patients 4 years old or younger, the Pediatric Glasgow Coma Scale is recommended (Table 19-2).

Certain numerical values of the Glasgow Coma Scale have been used to define the severity of TBI:

Both severe and mild to moderate TBI represent significant neurological diseases for children. Although mild to moderate traumatic brain injury very infrequently will result in severe disability or death, outcomes are variable and long-term sequelae have been reported.

Although our understanding of the mechanisms of damage to the immature brain after TBI remains quite limited, it is becoming clear that mechanisms of cellular injury and death in the immature brain are different from those described to date in the mature brain. Just as comprehensive evaluations of outcome describe functional differences, modern biomechanical and biochemical techniques have demonstrated that cellular and molecular responses to injury differ in the immature brain. That such differences exist seems intuitive given critical periods of vulnerability for the developing nervous system. As an example, better characterization of cerebral blood flow before and after injury is needed to enable appropriate strategies that will optimize oxygen delivery to the brain in children.

Pathologic changes that occur following trauma may be considered to occur in two stages. Damage that occurs at the time of the injury is referred to as the primary injury. This includes damage to axons, neurons, and cerebral blood vessels that occurs as a result of the force of the injury—that is, shear and strain forces that result in tearing of delicate cerebral structures. This injury may result in diffuse axonal injury, intracerebral/ subdural/epidural hematomas (from disruption of cerebral or dural blood vessels), or cortical contusions. Injury to the brain that occurs at any time after the initial impact of the traumatic event is termed the secondary injury. Events that may result in secondary injury to the brain and that have been shown to adversely affect outcome include hypoxia, hypotension, and hyperthermia. These types of injuries are preventable by caregivers, and their prevention may result in significant improvements in outcome. Figure 19-1 illustrates the concept of primary and secondary injury. Secondary injury may irreversibly damage brain tissue that was weakened by, but survived, the primary injury.

Mild and Moderate Traumatic Brain Injury

Mild head injury is a term first described by Rimel and associates and was taken to mean GCS 13 to 15 in a 1981 article.12 Although minor, head injuries not associated with a loss of consciousness constitute one of the most common public health problems. A certain number of children with minor head injuries (GCS 14-15) who have also lost consciousness are found to have significant intracranial pathology requiring neurosurgical procedures. With that in mind, a number of authors have attempted to redefine mild TBI since that original article. Because of the immense magnitude of the problem, numerous organizations, including the American Academy of Pediatrics, the American Academy of Family Physicians, and the American Academy of Neurology, have come up with practice parameters that address different aspects of mild pediatric head injury.13,14

Concussion is an entity that is commonly seen in pediatric patients with mild to moderate head injury. It is defined as a trauma-induced alteration in mental status that may or may not involve loss of consciousness.14 Patients with a concussion experience an immediate loss of consciousness, suppression of reflexes, transient cessation of breathing, and amnesia. The reflexes return and the patient begins to reconnect with the surrounding environment. The time to recovery is variable and may take several hours to days. The duration of the amnestic period, particularly anterograde amnesia, is probably the most reliable indicator of the presence of pathologic lesions. Postconcussive emesis in children is another worrisome problem in the acute setting. It can affect nearly one-half of patients and usually resolves within a few hours. Postconcussive emesis has not been associated with an increasing frequency of intracranial lesions.15 Concussions also occur without a defined loss of consciousness. Sports-related concussions are another common entity that primary care physicians, pediatric neurologists, and pediatric neurosurgeons have to deal with every day. Most children will recover from a sports-related concussion though recovery may take time, up to several months after the injury. The symptoms following a concussion are divided into three categories. Somatic complaints include headaches, fatigue, sleep disturbance, balance problems, and sensitivity to light and noise. Emotional and behavioral problems can manifest as irritability, lowered frustration tolerance, increased emotionality, depression, anxiety, and other personality changes. The last, and some feel the most important, category, especially in the school-age child is cognitive problems. Children may have slowed thinking, poor concentration, distractibility, trouble with learning and memory, and other problem-solving difficulties.

Severe Traumatic Brain Injury

Severe TBI in children carries a mortality of 8% to 22%.16,17 Approximately seven children between the ages of 0 to 14 years die every day, and many more suffer permanent disabilities as a result of TBI.4

Severe TBI is defined as a GCS score of 8 or less. By definition, children with severe TBI present with significant abnormalities in their neurological exam including coma, abnormal breathing (including at times apnea), focal neurological deficits, seizures, and even signs of herniation. Evolving brain edema or intracranial bleeding can result in acute deterioration of the child's neurological exam. Consideration should always be given to extracranial injuries, since such injuries can result in secondary injury to the brain.

Mild and Moderate Traumatic Brain Injury

In the pediatric head injury population, the most common clinical situation will be the child who rapidly regains consciousness after an injury. By the time the child is examined, the child will commonly have a GCS score of 13 to 15 or a mild head injury. The child also has a defined risk for having an intracranial injury. The Glasgow Coma Scale has long been considered a reliable indicator of severity of injury.

A number of tools have been developed to aid in the assessment of concussion. One of the most commonly used tools is the Standard Assessment of Concussion. It has been extensively validated and has significant normative data. The concussion symptom index has also been used, especially in athletes with normative data in high school as well as college athletes.18 Additionally, neuropsychological testing is one of the most important tools for detecting deficits in pediatric TBI, especially in patients with mild TBI. In sports-related concussion, the baseline use of neuropsychological testing has been used to determine return-to-play parameters in the absence of postconcussive symptoms.

Skull radiographs have a limited role in the diagnosis of intracranial injury following either mild or moderate pediatric head injury. Given the low predictive value of skull radiographs, the American Academy of Pediatrics recommends that cranial computed tomography (CT) scanning is the desirable imaging modality.13 Over 5 million children present to emergency rooms across the country with this diagnosis, and CT scanning of each of these patients, while considered the gold standard, would overwhelm the facilities and expose countless children to pointless radiation. Because of the higher degree of intracranial injury associated with a GCS score of less than 15, the authors recommend that any patient with a GCS less than 15 undergo a head CT. Another potential high-risk category is the patient who presents with a focal neurological deficit. The frequency of an intracranial injury in this population has been reported to be 11%.19

While the utility of computed tomography after minor head injury has been extensively studied, and different clinical criteria for predicting intracranial injury have been identified (Table 19-3), CT scanning can be associated with increased risk for long-term cancer, the need for sedation to accomplish the scan, and a prolonged ED evaluation, and this in turn has lead to multiple studies attempting to identify clinical predictors of intracranial injury. Boran and associates studied 421 patients with a GCS of 15 and were able to determine injuries found by either CT or plain radiographs. Intracranial lesions were found in 8.8% of the patients. They were able to determine that the only clinical parameters associated with an increase in intracranial lesion were posttraumatic seizures and loss of consciousness.15 In NEXUS (National Emergency X-Radiography Utilization Study), a large study looking at 1666 pediatric blunt trauma patients, CT was found to detect significant intracranial injury in 138, or 8.3%, of the patients. These authors also attempted to utilize a modification of the University of California-Davis Pediatric Head Injury Rule to identify patients with an intracranial injury. The sensitivity of their tool was found to be only 90.4%, with 13 children being misclassified as low risk.

Posttraumatic sequelae, especially in those patients with mild head injury, are significant, with upward of 15% of patients experiencing disabling symptoms up to 1 year after injury. Many of these patients do not have abnormalities on CT imaging. Limitations of CT scanning in detecting injury has lead to the use of magnetic resonance imaging (MRI) to detect diffuse axonal injury (DAI) and deliver prognostic information in cases of mild to moderate TBI. DAI is a type of injury characterized by significant axonal damage or shearing in brain regions such as the corpus collosum, parasagittal white matter, and gray-white matter junctions. Diffusion tensor imaging measures the microscopic random translational motion of water molecules within tissue. Wilde and others studied children with DTI and uninjured controls, and their data suggested that DTI may serve as an indicator for predicting outcome in pediatric patients with TBI.25 Magnetic resonance spectroscopy (MRS) is another imaging modality that is currently being used at some centers in the evaluation of patients with TBI. MRS has been shown to detect biochemical alterations, yet its value in providing truly valuable prognostic information in patients with mild to moderate TBI remains to be determined.

Severe Traumatic Brain Injury

Motor vehicle accidents and falls constitute the most common causes of severe TBI. In infants, the most common cause of severe TBI is child abuse, and a higher index of suspicion for intracranial injury may be required owing to an inconsistent clinical history.

The diagnosis of severe TBI is usually made based on the history and initial clinical findings. Once the diagnosis is made, children with severe TBI require close and continuous neurological monitoring, as their condition may rapidly deteriorate. In addition to abnormal neurological findings, including a GCS score of 8 or less, children with severe TBI can have signs of skull fractures, such as raccoon eyes, retro-orbital hematomas, and palpable abnormalities of the skull. The initial clinical diagnosis is usually confirmed by a CT scan in the acute phase, although children with severe brain injury can present with an initially normal brain CT. The initial scan is important because it may identify injuries that are amenable to surgical treatment, such as intracranial hematomas. Patients with a history of severe TBI and an abnormal CT scan are at increased risk of developing increased intracranial pressure.

While CT continues to be the primary diagnostic imaging modality in severe TBI, advanced MRI techniques, including diffusion tensor imaging and MR spectroscopy, are promising techniques with increased sensitivity, better prognostic capabilities, and also insights into the pathophysiology of brain injury after trauma.

Shaken Baby Syndrome

Shaken baby syndrome (SBS) is an important category of TBI in children, and children victims of SBS can present with signs and symptoms of severe TBI. In fact, SBS is the most common cause of traumatic brain injury in infants. As an entity, SBS provokes significant medical legal issues and requires careful attention to the etiology of the injury. Patients with SBS are difficult to evaluate because they often present with symptoms only, and if trauma is reported, it is usually minor. A patient's clinical presentation can vary significantly depending on the severity of the injury. Symptoms can range from mild lethargy and seizures to coma and even death.

Computed tomography is usually the initial diagnostic study. CT findings include subdural as well as subarachnoid hemorrhage. These are often thin, exerting no mass effect, but cover the entire hemisphere. In cases where the injury is severe, patients can present with diffuse loss of gray-white differentiation on CT scan, reflecting widespread cortical loss.26 Once the child has been stabilized and is not felt to be in a life-threatening situation, most major children's hospitals have a child protection service responsible for the complete medical workup of this patient. Workup includes a skeletal survey to assess for the presence of new or old fractures. Ophthalmology also must assess the child for the presence of retinal hemorrhages. Disorders such as coagulopathies and metabolic disorders need to be ruled out.

Mild to Moderate Traumatic Brain Injury

Initial evaluation of head-injured patients should always be performed according to the trauma protocols, with assessment of the ABCs and a focused history and physical exam. Life-threatening injuries should be assessed and supersede all other concerns. Once the child is deemed stable, then especially in patients with mild head injuries, a detailed physical exam is important. Concussion is probably one of the most important entities in pediatric TBI. Clinicians generally accept that no two concussions are exactly alike, making it necessary to individualize care within the parameters of their situation and rational clinical practice. Although no specific therapies for concussion-related symptoms currently exist, a number of experts have provided guidelines on the management of concussions as related to timing for return to play (Table 19-4).27

The overall incidence of posttraumatic seizure ranges between 5.5% and 21%. Seizures are usually classified based on the time to occurrence. The incidence of posttraumatic seizure also varies with decreasing GCS, as one might expect. Children with mild TBI have a seizure risk of approximately 2% to 6%.This increases to 12% to 27% in patients with moderate TBI, and the incidence of posttraumatic seizures also increases with younger age. The interval between the head injury and the first seizure can vary. Immediate seizures have a better prognosis in determining whether or not someone will be at risk for developing posttraumatic epilepsy. In contrast, the occurrence of a single late posttraumatic seizure has been used to define posttraumatic epilepsy. Inflicted (SBS) versus noninflicted TBI patients also differ in the rates of posttraumatic epilepsy. Victims of SBS have a rate of 48% to 65%, much higher than the 15% to 17% seen in patients after noninflicted trauma. This is most likely owing to the higher incidence of subdural hematomas in patients with SBS.28

Severe Traumatic Brain Injury

Treatment of patients with severe traumatic brain injury (GCS score of 8 or less) remains controversial. Much remains to be learned about the pathophysiology of brain injury and strategies for brain resuscitation in children. Although implementation of trauma programs and access to pediatric intensive care have improved outcome, there is only limited evidence behind treatment guidelines for children suffering severe TBI and there are no specific pharmacologic therapies available. Current treatment strategies are largely supportive.

Nevertheless, significant progress has been made over the past two decades. Despite the lack of class I evidence that demonstrates long-term, significant efficacy for any given treatment modality or intervention, mortality and morbidity from TBI have improved for both children and adults. Guidelines for in-hospital treatment of severe head injury in adults were published in 1997, and a revision of such document became recently available. Guidelines for prehospital management of traumatic brain injury in adults and children were also published in 2000 by the Brain Trauma Foundation (www.braintrauma.org). Finally, in July 2003 the Guidelines for the Acute Medical Management of Severe Traumatic Brain Injury in Infants, Children and Adolescents were published.29 The purpose of these documents was to upgrade practice parameters from opinion to guidelines that are supported by the best evidence available. It is important to note that most recommendations in this document are at the "option" level, given the lack of clinical evidence to support guideline and standard-of-care level recommendations. Upgrading practice parameters to standards of care—the strongest recommendation a guideline can make—will require new knowledge acquired through carefully planned clinical research.

Over the past years a significant amount of laboratory and human evidence has accumulated that demonstrates the impact of physiologic variables such as temperature, glucose level, and blood pressure on outcome. Protocols or guidelines focused in controlling such variables from the scene to the intensive care unit are likely to impact outcome. Also, standardized management of patients may improve the quality of clinical trials testing new therapies in pediatric head injury.

The main goal of current management protocols is to focus on normalizing physiologic variables that impact outcome after brain injury, both in children and adults. Therapeutic interventions known to decrease intracranial pressure (ICP) with a safe therapeutic profile are included. Second-tier therapies shown to be beneficial but also associated with significant side effects (ie, pentobarbital coma) are offered as treatment options for cases of persistent intracranial hypertension.

Management Protocol

Clinical judgment should be used to individualize patient management. As previously discussed, many aspects of care can only be considered treatment options. Treatment pathways apply to all patients admitted with severe TBI caused by

• Accidental trauma (blunt or penetrating trauma to the head)
• Gunshot wounds to the head
• Nonaccidental head trauma

Secondary Insult Prevention

The objective of intensive care management of severe traumatic brain injury is to prevent secondary insults to the traumatized brain. This may be achieved by direct and focused attention to those insults known to be associated with poor neurological outcome. This includes (1) prevention of cerebral ischemia and hypoxia and (2) prevention of cerebral hyperthermia.

Cerebral Ischemia

Cerebral ischemia may occur as a direct result of inadequate cerebral perfusion, decreased oxygen or glucose supply, increased metabolic demand of the brain, or from increased cerebral vascular resistance.

Cerebral Perfusion

Inadequate perfusion may have a direct and deleterious effect on the pathophysiology of brain injury. Cerebral perfusion is described in terms of cerebral perfusion pressure (CPP). CPP is calculated as the difference between mean arterial pressure (MAP) and mean intracranial pressure (ICP). That is, CPP = MAP − ICP. Current recommended guidelines suggest that CPP should be maintained around 60 mm Hg in adult patients. CPP values should be interpreted in the context of other perfusion indexes if those are available.

• Lower threshold levels of 50 mm Hg (2-6 years), 55 mm Hg (7-10 years), and 60 mm Hg (11-16 years) have been recommended for children. For adolescents, a minimum CPP of 60 mm Hg is recommended. Patients with CPP below these thresholds can experience a worse outcome.
• Optimal CPP levels for children lesser than 2 years of age have not been established, but several studies demonstrate that a CPP of 40 mm Hg or less is associated with higher mortality and worse outcome in children of any age.30 We therefore consider 45 mm Hg a critical threshold for CPP in children less than 2 years of age.

Mean Arterial Pressure

Mean arterial pressure (MAP) is an important determining factor in the CPP equation. One of the difficulties we face in caring for the patient with a severe head injury is determining an appropriate MAP. Factors to consider include patient age31 and preexisting hypertension. Age-appropriate MAP (Table 19-5) should be maintained and adjustments of MAP to maintain CPP above critical thresholds may be necessary.

It is also important to note that early systolic hypotension has also been associated with poor outcome after TBI in multiple pediatric clinical studies. Different definitions of hypotension have been used. A recent study described an association between early hypotension and outcome using the 75th percentile for age-appropriate systolic blood pressure (AASBP) as the threshold for hypotension. Maintaining AASBP from the initial stabilization phase is important and may influence outcome in children with traumatic brain injury.32 Hypotension is harmful after pediatric neurotrauma and should be avoided with appropriate volume resuscitation and vasopressor support.

Intracranial Pressure

The skull is a rigid structure containing cerebrospinal fluid (CSF), blood, and brain tissue. After TBI the normal balance between these intracranial contents may be disturbed, resulting in increasing pressure within the skull. Importantly, even infants and neonates with open fontanels can develop increased ICP once the intracranial volume expansion overwhelms the compliance of the immature skull and meninges. Elevated ICP has been shown to have definite prognostic implications in children with severe brain injury. In addition, it is generally held that treatment of elevated ICP may improve outcome in the patient with severe TBI. Measurement of ICP enables (1) calculation of CPP; (2) monitoring progress; and (3) assessment of treatment effectiveness. Indications for ICP monitoring in children with severe TBI include

• Patients with a GCS of 3 to 8 after initial resuscitation, who have an abnormal CT scan on admission (ie, hematoma, contusion, compressed basal cisterns, or edema).
• The treating physician may choose to monitor ICP in certain patients for whom serial neurological examination is precluded by sedation/analgesia or anesthesia. ICP monitoring is appropriate in children with severe TBI and a normal CT scan if motor posturing is noted on admission. ICP monitoring should be considered in children with severe TBI and a normal CT scan for those who are noted to have hypotension on admission and in children whose neurological exam does not improve or deteriorates. The presence of open fontanels and/or sutures in an infant with severe TBI does not preclude the development of intracranial hypertension or negate the utility of ICP monitoring.

The following methods are commonly used to monitor ICP in children with severe traumatic brain injury:

• Ventriculostomy. Placement of a ventriculostomy catheter and transduction of the fluid column is the mainstay of ICP monitoring. Placement of a catheter within the ventricle also offers a treatment option for raised ICP by allowing CSF drainage. When measuring the pressure, it is necessary to always close off the drainage system prior to recording the pressure readings. Problems with monitoring ICP by ventriculostomy and a transduced fluid column include difficulty with insertion, particularly if the ventricles are small; variable accuracy, because fluid-filled columns are subject to poor waveform, especially if the ventricles are small or collapsed; mechanical problems such as air bubbles, blood clots, or debris in tubing; and infection.
• Ventriculostomy with built-in microsensor. New-generation ventriculostomy catheters do not require transduction of a fluid column. They have a built-in miniaturized strain gauge transducer in the tip of the catheter, which gives continuous accurate ICP readings even while simultaneously/continuosly draining CSF.
• Intraparenchymal bolt. This may be the ideal method of monitoring ICP if the ventricular system is not accessible. A pressure sensor is introduced into the brain parenchyma and held in place by a bolt secured into the skull.

ICP Management

Normal adult ICP values range from 5 to 15 mm Hg. Normal ICP values tend to be lower in infants and younger children (Table 19-4). Increased ICP in children with severe TBI correlates with worse outcome and also decreased cerebral blood flow.33 Careful management of the patient with ICP monitoring is important, as good ICP control may improve outcome. Basic measures aimed at maintaining adequate venous drainage from the head are safe and simple, and promote normal ICP. Overly tight cervical collars as well as circumferential taping of the endotracheal tube can restrict venous drainage from the head, which may increase ICP. Cervical flexion or extension and lateral head rotation can alter cerebral venous outflow. Maintaining neutral head alignment promotes venous drainage. As demonstrated in Figure 19-2, the effect of cerebral venous outflow on ICP can be significant. Elevating the head of the bed (HOB) 30 degrees by flexing the patient at the hips—not at the abdomen—will also promote venous drainage. However, there may be instances when raising the HOB is contraindicated. If CPP is low, it will be preferable to position the patient in a supine position to increase CPP. To avoid orthostatic hypotension, HOB elevation should be considered with euvolemic patients only. Additionally, in patients with a spine not cleared, the reverse Trendelenburg position can be used to elevate the head without disrupting the spinal alignment.

The currently recommended threshold for treatment of elevated intracranial pressure is 20 mm Hg. In small children, 15 mm Hg might be more appropriate, but data supporting this lower threshold are lacking. Interpretation and treatment of intracranial pressure based on this threshold value should be made in conjunction with careful attention to maintaining CPP above critical values.

There are many causes of raised ICP that are not directly related to intracranial pathology. These factors should be excluded and/or treated prior to initiating other treatments. These factors include

1. Hyperthermia

2. Hypoxemia

3. Agitation and pain

4. Shivering

5. Resistance to mechanical ventilation

6. Seizures

For children with persistent intracranial hypertension, the methods described in the following sections may result in substantial reductions in ICP, improved cerebral perfusion, and may improve clinical outcome.

Sedation and Analgesia

If patients are inadequately sedated the sympathetic nervous system is overactive. This may result in significant increases in intracranial blood volume. After TBI, the brain's ability to compensate for this increase in blood volume is impaired and the intracranial pressure will rise. It is important to ensure that adequate sedation and analgesia are given to avoid this situation. We recommend midazolam and fentanyl as first-line therapy. If the level of sedation and analgesia is not optimal, morphine may be substituted for fentanyl, but should be used carefully in patients with abnormal renal function. Sedatives and analgesics should be carefully titrated to avoid hemodynamic instability. In compliance with current published recommendations from the FDA, continuous infusions of propofol should not be used for continuous sedation in the pediatric intensive care unit. Exceptions for short periods of sedation (ie, less than 6 hours) can be made at the discretion of the neurocritical care team and require appropriate hemodynamic and biochemical monitoring for indicators of propofol side effects (ie, hypotension, bradycardia, lactic acidosis, lipemia, evidence of rhabdomyolysis).

Neuromuscular Blockade

Neuromuscular blockade (NMB) can potentially reduce mean airway pressure, which can facilitate cerebral venous outflow, prevent shivering or posturing, and reduce metabolic demands. In contrast, NMB carries the risk of masking seizures, cardiovascular side effects, and immobilization stress (if inadequate sedation and analgesia); increases the risk of pneumonia in adult patients with TBI; and increases length of stay in the ICU.34 After careful consideration of the potential benefit and risk, NMB can be considered as second-tier therapy for persistent intracranial hypertension (ICP > 20 mm Hg) and/or perfusion deficits (low CPP or brain tissue oxygen tension). Since there are no validated tools to monitor NMB in comatose patients with acute brain injury, we recommend that NMB agents be discontinued periodically, for example, every 24 hours. If no benefit is noted, NMB can be discontinued. If the patient becomes difficult to ventilate or ICP increases, NMB can be reinstituted and evaluated again (Figure 19-3). Considering the potential risk associated with the use of NMB, we recommend avoiding this strategy if no benefit on ICP or the ability to mechanically ventilate can be demonstrated.

Cerebrospinal Fluid (CSF) Drainage

If a ventriculostomy is present, CSF may be drained intermittently or continuously. The reservoir is placed above the external auditory canal at a predetermined height. Overdrainage of CSF should be avoided as it could collapse the ventricles. Lumbar puncture should not be done in acute TBI owing to risk of herniation. The addition of lumbar drainage can be considered as an option only in the case of refractory intracranial hypertension with a functioning ventriculostomy, open basal cisterns, and no evidence of a major mass lesion or shift on imaging studies. In pediatric patients with refractory intracranial hypertension, controlled lumbar drainage has been successful in lowering ICP.35

Hyperosmolar Therapy

Mannitol is an osmotic diuretic and an effective treatment for patients with increased ICP caused by cerebral edema. Mannitol decreases ICP by several proposed mechanisms, including shifting fluid from the brain to the intravascular space and decreasing CSF production. Mannitol may also improve cerebral blood flow by reducing blood viscosity. When treating patients with mannitol, hypotension caused by dehydration and hypovolemia should be avoided. Osmolar gap measurements correlate with mannitol levels better than serum osmolarity and should be maintained below 20.36,37 Using osmolar gap calculations also avoids confusion when using mannitol and hypertonic saline concurrently. Hypertonic saline is effective to control increased ICP after severe head injury. A continuous infusion administered on a sliding scale (the minimum dose required to control ICP should be used) is recommended (Figure 19-4). Serum sodium of 150 to 160 mEq/L, and a serum osmolality level of 360 mOsm/L, appear to be tolerated with hypertonic saline, although cases of renal failure have been reported.38 Until more safety and efficacy information becomes available, serum sodium levels greater than 165 mEq/L should be avoided. Rapid changes in serum sodium in children with hypernatremia can result in worsening intracranial hypertension. The long-standing acceptance, safety, and limited evidentiary support of the efficacy of mannitol therapy should be weighted against the limited clinical experience but reasonably good performance of hypertonic saline.

Hyperventilation

In the past, hyperventilation was often employed as a means to reduce high ICP. Current practice now supports avoiding hyperventilation unless signs of herniation are exhibited. Hyperventilation lowers ICP by causing vasoconstriction and decreasing cerebral blood flow (CBF). For every 1 mm decrease in PCO2 there is a 3% decrease in CBF. Accordingly, PCO2 should be kept at 35 to 40 mm Hg. The caregiver should anticipate the need for ETCO2 monitoring. In the setting of hyperventilation as a second-tier therapy to treat refractory intracranial hypertension, perfusion indexes such as brain tissue PO2, jugular bulb hemoglobin oxygen saturation, and/or cerebral blood flow (CBF) should be monitored. Although hyperventilation-induced cerebral ischemia remains a controversial topic, the use of prophylactic aggressive hyperventilation has been associated with worse outcome in adult patients with TBI.39

Cerebral Oxygenation

Cerebral hypoxia can exacerbate the deleterious effects of acute traumatic brain injury. Despite aggressive treatment for ICP and CPP, cerebral hypoxia can be persistent. Depth and duration of brain tissue hypoxia has been associated with increased mortality in severe TBI patients.40

We now have the ability to measure brain tissue oxygen pressure (PbtO2) continuously by a small probe placed into brain tissue and secured to the skull by a bolt. The PbtO2 monitor measures intraparenchimal oxygen and temperature and is intended to be used in conjunction with other monitors to help indicate the perfusion status of cerebral tissue. It is intended to provide data additional to that obtained by current clinical practice in cases where ischemia or hypoxia is a concern. Indications for PbtO2 monitoring are the same as for ICP monitoring and include any patient with a GCS of 3 to 8, an abnormal CT scan of the brain, and/or abnormal clinical exam. The use of both ICP and PbtO2monitors and therapy directed at PbtO2 is associated with reduced patient death following severe TBI.41 These associations are also supported by neuropathology and physiology preclinical data on the consequences of low PbtO2 values.

Normal human brain has a critical PbtO2 between 15 and 20 mm Hg, below which infarction of tissue may occur. In clinical practice, a target PbtO2 of 25 to 35 mm Hg has been adopted. Table 19-6 outlines some of the underlying factors that may be responsible for abnormal values. PbtO2 values must be interpreted in conjunction with other monitoring parameters and the clinical situation.

The most common clinical scenario will be a low PbtO2 (<25 mm Hg). It is important to ensure that the monitor is functioning normally and that the probe is correctly connected and calibrated (Figure 19-5). Consideration can be given to placing patient on 100% FiO2 for 15 minutes as a temporary intervention while underlying causes are investigated. The main priority is to optimize other parameters in order to decrease FiO2 and increase PbtO2: Is the patient adequately sedated? Is the patient hydrated and hemodinamically stable? Does the patient have significant anemia? Is the patient hyperthermic? Is the patient being hyperventilated? There may be other parameters that are outside the goal of management range, such as high ICP. Additional management suggestions are shown in the severe TBI management flow sheet (Figure 19-6). After considering and attempting to correct all relevant parameters, attempts can be made to correct critical PbtO2 values by titrating the FiO2 to keep the patient's PbtO2 lesser than 20 mm Hg. When titrating FiO2 for prolonged periods of therapy, it is suggested that the FiO2 be kept greater than or equal to 60%. In the context of low brain oxygenation, glucose control (serum glucose 80-180) is important after TBI because of the deleterious effects of hyperglycemia or hypoglycemia in the setting of brain ischemia.

Cerebral Hyperthermia

Elevated temperature can contribute to worsening brain injury after severe TBI. A number of trials are looking at hypothermia as a management technique to improve outcome after traumatic brain injury. It has been reported to be effective in some centers, but at present this treatment has not been proven to be effective in a multicenter trial. While the efficacy of hypothermia after severe TBI in children remains to be determined, our management approach does include temperature control aimed at maintaining normothermia. Brain temperature can now be measured continuously. The brain temperature probe comes as part of the PbtO2 setup and does not require any additional equipment. Our current management strategy is to maintain a brain temperature of 37 ± 0.5°C. This we achieve by active cooling as required. If brain temperature monitoring is not available, continuous temperature monitoring using either an esophageal, rectal, or Foley catheter temperature probe is recommended. Oral and axillary temperature measurements are discouraged, especially when using external cooling. Fever is combated aggressively because of the impact that high temperature can have in outcome after several types of brain injury, including trauma. While attempting to normalize brain/core temperature, shivering remains an issue. Shivering is considered a problem not only because it is uncomfortable to patients but also because it can increase ICP. Pharmacologic and nonpharmacologic interventions to avoid shivering should be attempted in order to assure effective temperature regulation and prevent raises in ICP. Such interventions may include transient use of neuromuscular blockade.

A Coordinated Target-Specific Management Approach

The primary concern when taking care of children with severe TBI is maintaining adequate cerebral perfusion—"perfuse it or lose it"—while optimizing relevant physiologic and clinical variables such as temperature and level of sedation. The initial monitoring setup may include an arterial line and central venous line if hemodynamically indicated; ICP monitor—ventriculostomy or intraparenchymal microsensor; pulse oximetry; end- tidal CO2, and the PbtO2 and temperature probes. The risk/benefit ratio of all monitoring devices and therapeutic interventions should be carefully analyzed by the multidisciplinary team usually involved in the complex care of this fragile patient population.